To start out, only actively-synced is returned.
For example, this is useful for jobs that started out in background
mode and switched to active mode. Once actively-synced is true, it's
clear that the mode switch has been completed. Note that completion of
the switch might happen much earlier, e.g. if the switch happens
before the job is ready, once all background operations have finished.
It's assumed that whether the disks are actively-synced or not is more
interesting than whether the mode switch completed. That information
can still be added if required in the future.
In presence of an iothread, the actively_synced member is now shared
between the iothread and the main thread, so turn accesses to it
atomic.
Requires to adapt the output for iotest 109.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231031135431.393137-10-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In preparation to turn BlockJobInfo into a union with @type as the
discriminator. That requires it to be an enum. Even without that
requirement, it's nicer to have an enum instead of a str here.
No functional change is intended.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-ID: <20231031135431.393137-7-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
which allows switching the @copy-mode from 'background' to
'write-blocking'.
This is useful for management applications, so they can start out in
background mode to avoid limiting guest write speed and switch to
active mode when certain criteria are fulfilled.
In presence of an iothread, the copy_mode member is now shared between
the iothread and the main thread, so turn accesses to it atomic.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231031135431.393137-6-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In preparation to allow changing the copy_mode via QMP. When running
in an iothread, it could be that copy_mode is changed from the main
thread in between reading copy_mode in bdrv_mirror_top_pwritev() and
reading copy_mode in bdrv_mirror_top_do_write(), so they might end up
disagreeing about whether copy_to_target is true or false. Avoid that
scenario by determining copy_to_target only once and passing it to
bdrv_mirror_top_do_write() as an argument.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-ID: <20231031135431.393137-5-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In preparation to allow switching to active mode without draining.
Initialization of the bitmap in mirror_dirty_init() still happens with
the original/backing BlockDriverState, which should be fine, because
the mirror top has the same length.
Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231031135431.393137-4-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
There is a batching mechanism for virtio-blk Used Buffer Notifications
that is no longer needed because the previous commit added batching to
virtio_notify_irqfd().
Note that this mechanism was rarely used in practice because it is only
enabled when EVENT_IDX is not negotiated by the driver. Modern drivers
enable EVENT_IDX.
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230913200045.1024233-5-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
virtio-blk and virtio-scsi invoke virtio_irqfd_notify() to send Used
Buffer Notifications from an IOThread. This involves an eventfd
write(2) syscall. Calling this repeatedly when completing multiple I/O
requests in a row is wasteful.
Use the defer_call() API to batch together virtio_irqfd_notify() calls
made during thread pool (aio=threads), Linux AIO (aio=native), and
io_uring (aio=io_uring) completion processing.
Behavior is unchanged for emulated devices that do not use
defer_call_begin()/defer_call_end() since defer_call() immediately
invokes the callback when called outside a
defer_call_begin()/defer_call_end() region.
fio rw=randread bs=4k iodepth=64 numjobs=8 IOPS increases by ~9% with a
single IOThread and 8 vCPUs. iodepth=1 decreases by ~1% but this could
be noise. Detailed performance data and configuration specifics are
available here:
https://gitlab.com/stefanha/virt-playbooks/-/tree/blk_io_plug-irqfd
This duplicates the BH that virtio-blk uses for batching. The next
commit will remove it.
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20230913200045.1024233-4-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The bdrv_getlength() function is a generated co-wrapper and uses
AIO_WAIT_WHILE() to wait for the spawned coroutine. AIO_WAIT_WHILE()
expects the lock to be acquired exactly once.
Fix a case where it may be acquired twice. This can happen when the
source node is explicitly specified as the @replaces parameter or if the
source node is a filter node.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231019131936.414246-4-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
by passing the BlockDriverState along, so the held AioContext can be
dropped before polling. See commit 31b2ddfea3 ("graph-lock: Unlock the
AioContext while polling") which introduced this functionality for
more information.
The only way to reach bdrv_close() is via bdrv_unref() and for calling
that the BlockDriverState's AioContext lock is supposed to be held.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231019131936.414246-3-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Same rationale as in 31b2ddfea3 ("graph-lock: Unlock the AioContext
while polling"). Otherwise, a deadlock can happen.
The alternative would be to pass a BlockDriverState along to
bdrv_graph_wrlock(), but there is no BlockDriverState readily
available and it's also better conceptually, because the lock is held
for the job.
The function is always called with the job's AioContext lock held, via
one of the .abort, .clean, .free or .prepare job driver functions.
Thus, it's safe to drop it.
While mirror_exit_common() does hold a second AioContext lock while
calling block_job_remove_all_bdrv(), that is for the main thread's
AioContext and does not need to be dropped (bdrv_graph_wrlock(bs) also
skips dropping the lock if bdrv_get_aio_context(bs) ==
qemu_get_aio_context()).
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
Message-ID: <20231019131936.414246-2-f.ebner@proxmox.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
iotests case 118 already tests all relevant operations for media change
with multiple devices, however never with iothreads. This changes the
test so that the virtio-scsi tests run with an iothread.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20231013153302.39234-3-kwolf@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
blk_insert_bs() requires that the caller holds the AioContext lock for
the node to be inserted. Since commit c066e808e1, neglecting to do so
causes a crash when the child has to be moved to a different AioContext
to attach it to the BlockBackend.
This fixes qmp_blockdev_insert_anon_medium(), which is called for the
QMP commands 'blockdev-insert-medium' and 'blockdev-change-medium', to
correctly take the lock.
Cc: qemu-stable@nongnu.org
Fixes: https://issues.redhat.com/browse/RHEL-3922
Fixes: c066e808e1
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20231013153302.39234-2-kwolf@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The test cases considered so far:
314 (new test suite):
1. Check that compression mode isn't compatible with "-f raw" (raw
format doesn't support compression).
2. Check that rebasing an image onto no backing file preserves the data
and writes the copied clusters actually compressed.
3. Same as 2, but with a raw backing file (i.e. the clusters copied from the
backing are originally uncompressed -- we check they end up compressed
after being merged).
4. Remove a single delta from a backing chain, perform the same checks
as in 2.
5. Check that even when backing and overlay are initially uncompressed,
copied clusters end up compressed when rebase with compression is
performed.
271:
1. Check that when target image has subclusters, rebase with compression
will make an entire cluster containing the written subcluster
compressed.
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-9-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
If we rebase an image whose backing file has compressed clusters, we
might end up wasting disk space since the copied clusters are now
uncompressed. In order to have better control over this, let's add
"--compress" option to the "qemu-img rebase" command.
Note that this option affects only the clusters which are actually being
copied from the original backing file. The clusters which were
uncompressed in the target image will remain so.
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-8-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
As the previous commit changes the logic of "qemu-img rebase" (it's using
write alignment now), let's add a couple more test cases which would
ensure it works correctly. In particular, the following scenarios:
024: add test case for rebase within one backing chain when the overlay
cluster size > backings cluster size;
271: add test case for rebase images that contain subclusters. Check
that no extra allocations are being made.
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-7-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
When rebasing an image from one backing file to another, we need to
compare data from old and new backings. If the diff between that data
happens to be unaligned to the target cluster size, we might end up
doing partial writes, which would lead to copy-on-write and additional IO.
Consider the following simple case (virtual_size == cluster_size == 64K):
base <-- inc1 <-- inc2
qemu-io -c "write -P 0xaa 0 32K" base.qcow2
qemu-io -c "write -P 0xcc 32K 32K" base.qcow2
qemu-io -c "write -P 0xbb 0 32K" inc1.qcow2
qemu-io -c "write -P 0xcc 32K 32K" inc1.qcow2
qemu-img rebase -f qcow2 -b base.qcow2 -F qcow2 inc2.qcow2
While doing rebase, we'll write a half of the cluster to inc2, and block
layer will have to read the 2nd half of the same cluster from the base image
inc1 while doing this write operation, although the whole cluster is already
read earlier to perform data comparison.
In order to avoid these unnecessary IO cycles, let's make sure every
write request is aligned to the overlay subcluster boundaries. Using
subcluster size is universal as for the images which don't have them
this size equals to the cluster size. so in any case we end up aligning
to the smallest unit of allocation.
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Message-ID: <20230919165804.439110-6-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add @chsize param to the function which, if non-zero, would represent
the chunk size to be used for comparison. If it's zero, then
BDRV_SECTOR_SIZE is used as default chunk size, which is the previous
behaviour.
In particular, we're going to use this param in img_rebase() to make the
write requests aligned to a predefined alignment value.
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20230919165804.439110-5-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Since commit bb1c05973c ("qemu-img: Use qemu_blockalign"), buffers for
the data read from the old and new backing files are aligned using
BlockDriverState (or BlockBackend later on) referring to the target image.
However, this isn't quite right, because buf_new is only being used for
reading from the new backing, while buf_old is being used for both reading
from the old backing and writing to the target. Let's take that into account
and use more appropriate values as alignments.
Signed-off-by: Andrey Drobyshev <andrey.drobyshev@virtuozzo.com>
Message-ID: <20230919165804.439110-4-andrey.drobyshev@virtuozzo.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Fiona EbnerandKevin Wolf
Stefan HajnocziandKevin Wolf
Andrey DrobyshevandKevin Wolf